The beam analysis algorithm for path planning for redundant manipulators

Abstract

An algorithm for path planning for redundant manipulators is presented in this paper. The algorithm uses harmonic potential fields defined globally in W-space (work space) for both path planning and obstacle avoidance. Although paths generated by harmonic potential fields are collision free for point robots, this is not always the case for manipulators when especially tight manocuvring is required. To enhance collision avoidance ability of redundant manipulators, the beam analysis algorithm is proposed. The algorithm sends beams along the path generated for point robots to determine virtual obstacle points where collision with obstacle is likely to occur. The potential field is then regenerated to include these virtual obstacle points. Besides, the interaction between manipulator links and the potential field is accomplished by the control points situated on only proximal ends of the links. The virtual obstacle points and selected control points allow the manipulator to achieve tight manoeuvring in W-spaces cluttered with many obstacles. The improvement in performance is also clearly indicated by a benchmark scheme that compares the algorithms by means of the complexity of the environment with respect to link lengths of redundant manipulators. Furthermore, the beam analysis algorithm readily produces safer paths for mobile robots, which does not suffer too far or too close problems. Examples are included to demonstrate these features of the algorithm. © 2004 Elsevier Ltd. All rights reserved.